| dc.description.abstract | In robotics research, ensuring that robotic systems accurately detect and avoid both static and moving objects is the main challenge to building robots more autonomous.This thesis presents a method of accurately obtaining th'eposition of an obstacle from a robot system by considering the angular orientation of the obstaclein an indoor unstructured or structured environment. The method used is an advanced sonar sensing method that overcomes the problem of conventional ultrasonicranging that provides the direction of the reflecting point. Conventional ultrasonic sensors do not provide the reflecting point accurately, therefore, inconvenient to use in detecting obstacles where accuracy in range measurement is required. The method described in this thesis was implemented and tested on a physical robot. The robot, named Polyrobot, has two ultrasonic transducers at extreme ends of the robot's front panel at a fixed distance. The robot used was fully autonomous with all power and 'processing onboard. The robot was programmed in wiring language (Arduino C version 11). All experimental data were gathered in unaltered indoor laboratory environment with both static and dynamic obstacles. The ultrasonic transducers were used to measure distance between the robot and obstacles, with the distance between the two transducers taken as the constant obstacle width. The range measurements from the two transducers were uploaded to the microcontroller board on the robot and displayed on personal computer monitor. Logic functions using comparison operators were developed based on the range measurements at different obstacle angular position with reference to robot centre and incorporated in the algorithm using the Arduino C version 11. The logic functions in the algorithm are responsible for decision making, enabling the robot to make decisions on whether to move forward, turn right, turn left, stop or reverse in order to avoid an obstacle. The data from the sonar sensors were 'used as the input variables. The output variables were the curvatures (turns) and velocity of the robot wheel motors (actuators) in order to avoid an obstacle. The developed algorithm is the artificial | en_US |
